Overview
What are the Research and Qualification Agents?
Research Agents and Qualification Agents automate the process of qualifying & researching the company according to your Ideal Customer Profile (ICP).
Qualification Agents collect information and apply specific criteria to determine if an account matches your Ideal Customer Profile (ICP). Think of it as gatekeeper. Only those companies that fit the criteria will be given to your sales team. Qualification agents also decide into which Vertical the company fits.
Research Agents collect and format information based on your instructions. Use it to find signals, specific contextual information for personalization, enrich data like number of employees.
When should I use it?
To let agents know the criteria that define what qualifies as your ICP
When you want to scale research
When you want to find signals about accounts
Before You Start
You must have Admin access.
Optional: use the Research Agent Sandbox to test and fine-tune agents.
Step-by-Step Guides
Create a Research Agent from our Templates
Go to the Agent Training Center >> Ecosystems >> Specific ecosystem
Click on "Add Research Agent."
Choose a template from the “All templates” tab. For example: “Key Products and Services”.
Templates come with pre-set agent names, answer formats, and instructions tested by the Evergrowth team.
(Optional) If you need, you can customize the agent settings:
Agent Name: Choose a clear, descriptive name. This will be used in question formulation and for generating instructions.
Instructions: Write clear, step-by-step instructions for the agent. Specify exactly what information to find and how to format the answer (e.g., Yes/No, list, explanation).
Will be used in: Decide if the agent’s output should impact the business model description, account planning, or be marked as a signal for play generation.
Select LLM Model: Select the appropriate model (e.g., Sonar for broad web search, 4.1 for specific website navigation, Gemini for YouTube/video content).
Response type: Define the expected response type (text, number, select, multi-select, range, yes/no).
(Optional) Customize assignment to Vertical(s): Select which verticals this agent should be active in by turning on the toggle next to vertical name. By default, all Verticals are selected.
Click “Add” to add the agent. The agent will now appear at the bottom of the list and be available for use in workflows or in “Research agent”.
Create a Research Agent by using the “Generate with AI” function
Use this process for custom research needs not covered by templates.
Go to the Agent Training Center >> Ecosystems >> Specific ecosystem
Click on "Add Research Agent."
In the pop-up, click “Add new”.
Enter a clear, descriptive Agent Name that describes briefly what info you’re looking for. For example: “Has out of stock items?”
The name is used in qualification logic and for generating instructions with AI.
Click “Generate Instructions with AI"
(Optional) add additional context for what you’re looking for. Usually not needed.
Click “Generate”
The Agent will write the instructions and automatically choose the best suitable LLM model.
Assign the agent to the relevant vertical(s).
Click "Add" to save the agent.
Create a Qualification Agent
Qualification agents are similar to research agents but include criteria to define the ICP fit. The key difference here is that you need to define the criteria next to each vertical.
Go to the Agent Training Center >> Ecosystems >> Specific ecosystem
Click on "Add Qualification Agent."
Follow the steps as in "Create a Research Agent" (see above: using a template or from scratch).
Assign the Qualification Agent to the appropriate vertical(s) & define qualification criteria that accounts must meet to be marked as a fit.
The agent will use these criteria to qualify or disqualify accounts.
Save the agent.
Reuse existing Qualification/Research agents or Import from Sandbox
In the Agent Training Center, choose to reuse an existing agent from another ecosystem or import it from the Research Agent Sandbox.
Select the agent you want to reuse/import.
Adjust settings or instructions as needed.
Assign to the relevant vertical(s) and click “Add”.
FAQ
What is the difference between a Research Agent and a Qualification Agent?
A Research Agent gathers and formats information based on your instructions. A Qualification Agent does the same but also applies specific criteria to determine if an account is a fit for your ICP and what Vertical does it fit.
What LLM model should I use for my agent?
Use "Sonar" for broad web searches (e.g., job postings across multiple sites).
Use "4.1" for tasks that require visiting specific URLs or extracting data from a particular page (e.g., company values from an About page).
Use "Gemini" for finding and summarizing YouTube videos or podcasts.
Use "4.0 Mini" for simple, direct tasks (e.g., checking if a company has a privacy policy).
Can I assign an agent to multiple verticals?
Yes, you can enable or disable each agent per vertical. This allows for granular control over which agents are active in which segments.
What happens if I do not assign an agent to any vertical?
The agent will be created but will not be used until it is assigned to at least one vertical.
Additional Notes & Tips
If you seek consistency, then include an answer format in your instructions.
Agents can only access publicly available information. They cannot log into pages or access paid content or databases.
The agent name is important for both qualification logic and for generating instructions with AI.
Marking an agent as a "signal" allows it to be used in Plays.
You can test agents in the Research Agent Sandbox before adding them to an ecosystem.
Review and update agent instructions regularly to keep them aligned with your evolving GTM strategy.